Version classiqueVersion mobile

A Complex Polity

Anne-Marie Peatrik

Part one. Puzzling generations: or, how the system really worked

Chapter two

Mastering the succession of generations

Texte intégral

1Before their radical transformation during the colonial period, the age-sets of the Tigania-Igembe were clearly evident to observers. All the travelers to the area remarked, when they did not suffer the misfortune of experiencing it directly, the difference between the warriors, who often displayed a threatening attitude, and the elders, more inclined to conciliation, with whom “blood brotherhood” relationships were forged. The first colonial administrators drew the same conclusions and made a point of reducing, or manipulating, the warriors’ strength and imposing negotiations on the elders who were ready to accept them, up to a certain point. However, even though they did show an interest in the phenomenon, the administrators were largely incapable of understanding the mechanisms at work. Certain recurrent disorders were punished by arrests or a ban on gatherings; when this tactic of repression was insufficient to restore calm, the mediation of the elders was of little help in finding solutions: they were careful not to explain certain principles, either because it was not in their interest or because they found it difficult to talk about them.

  • 1  Even the English word “set” is a source of misunderstanding. This polysemous word conveys various (...)

2As much as the idea of a cycle to be accomplished and a scale of age grades to climb are easy to grasp, the conception that these populations have of their own sets is more difficult to penetrate. The Tigania-Igembe have a varied and precise vocabulary to designate their different sorts of sets, but the informants give different and contradictory versions of their constitution. An elder who has progressed through the whole cycle will obviously have a different view of the sets than a young man of more limited horizons, but the picture painted by the women elders is surprising in that it reflects a more comprehensive perception of the whole. The models and opinions are also marked by usual conflicts between successive sets, intensified by the unprecedented power issues generated by the colonial conquest. Thorough comparison between the viewpoints of the different informants and the partial but valuable descriptions available in the literature and archives has allowed us to adjust these heterogeneous views and bring to light the underlying operating principles1.

3The sets will be described on the basis of the most easily observed divisions. Individuals are (were) successively or simultaneously included in various different units, and we must distinguish between temporary affiliation and definitive assignment. Sometimes we are dealing with categories (groups with imprecise contours and fluctuating identity) and sometimes with sets: an individual cannot belong to two sets at the same time, and it is these sets that possess the attributes governing the exercise of power. The search for principles of recruitment and underlying rules will lead us to examine details that a less mindful observer might consider incidental, and to gradually grasp the principles at work in what can be qualified as a system: the whole population is (was) distributed in sets connected to the age grades in such a way that the movement of one set alters the positions of all the others. This system, in which the sets successively hold authority and embody the principle of sovereignty, is the foundation of the polity of the old Tigania-Igembe, who thereby differ from neighboring populations—allies or enemies— with more or less similar systems.

1. A foundation course in the system

4From an early age, the children of each neighborhood developed the habit of gathering together in groups, which would subsequently become bands of adolescents. There is nothing exceptional about this phenomenon, but it takes on a singular character here because, far from dissolving with the onset of adulthood, an individual’s age group continues to surround him throughout his life. It first appears with the rites held to celebrate the newborn baby’s emergence into the light of day. Its importance, which grows with the passing years, is attested to by the terms (substantives and proper nouns) used to designate the age groups of children and youths.

5The very young boys, the “small non-initiates” living in their mothers’ house, formed age groups called kathingiriti or luungî; this latter word means “lung” and refers to the fact that this is the cut of meat they got when carcasses were shared out. Joining in these age groups was not obligatory, but when a boy did want to join—and this always happened in the end—he had to offer the bigger boys a metal ring (kibuthu), beads or a stick. Girls were recruited in the same way. With the onset of adolescence, the boy became a “big non-initiate” and had to sleep in the gaaru ya ijîîjî, the collective dormitory which can be found in every neighborhood: built by the warriors and fathers and identified by its own specific name, this dormitory represents the first gaaru of his life. Given that the boys have to sleep here until their initiation, the range of ages is wide and contains subcategories. The terms used for these subcategories vary somewhat from one place to another, but they are recurrent and reveal a common three-way division. In Kîgûcwa, the following categories of adolescents were used, starting with the youngest:

  • murerwa, the “infants”, from the verb kurerwa, “to be fed”;

  • kîlelemi [literal meaning not found];

  • gîtile, the last category comprising the îjîîjî biî îkûrû, the “oldest big non-initiates”.

6Although it is not formalized, this differentiation nevertheless represents the discovery and learning of a hierarchical mode predominant here, since it structures the whole population. The “infants” were the servants of the older boys and might even be treated as slaves (nkombo), according to the elderly: they were the ones who had to fetch water and wood and sweep up, and they were beaten if considered to have behaved badly. The younger ones also had to bring the older ones tobacco and gruel prepared by their mothers. During public events where the different classes took part in sessions of singing and dancing, the youngsters from neighboring territorial sections competed in dances (nguri) and various games; the older ones confronted each other in a particularly popular form of wrestling known as ukindara, in which the victor, kaindio, was praised in songs of the same name that celebrate his exploits. The sections prided themselves on their champions, but it was impossible for these non-initiates to take part in actual raids, just as it would be unthinkable for warriors to indulge in these activities, which only involve the non-initiates.

Ill. 8 — “Embe [Igembe] shepherds”

Ill. 8 — “Embe [Igembe] shepherds”

1892-1893 Rokeby Archives (etching in Chanler 1896: 250).

Ill. 9 — Boys

Ill. 9 — Boys

Kianjai section, 1893-1894 (photo by G. Kolb, in Neumann 1898: 65).

7The oldest non-initiates, the îjîîjî biî îkûrû, who rule over life in the gaaru, eventually ended up forming a clandestine council, the ncibi, described sarcastically as the “kiama of the big non-initiates”. It is clandestine because, no matter how old they are, these boys are nevertheless non-initiates, and in this case, the term kiama is a misuse of language used to denounce a misuse of power. Although it was in principle illegal, a ncibi was formed in every territorial section. The eldest boys secretly recruited into the ncibi those who had gone through every stage up to uiji bukuru. Of course, this all took place at night. In olden times, new recruits were not beaten as they would be later; they had to offer a bundle of miraa, a sack of sorghum or millet, elsewhere arrows or necklaces. In certain cases, the neophyte was terrorized by a “wild beast”, the kîrimo—actually one of the older boys dressed in skins and leaves and armed with thorns, who attacked and scratched him as he went into his burrow, only releasing his “prey” when the latter agreed to pay him something. The new recruits were given a name known only to them, and were assigned a “godfather”, following a process imitating the code of the official initiation. Levels called njaru (white), njiro (black) and lomo, indicating a return back in time, also existed within the ncibi; the candidate attained each level according to his skill in answering certain songs and riddles issued as challenges, or in interpreting notches carved into a wooden object, the gichiumû, which explicitly reproduced the human sex organs. The notoriety of the ncibi was further tarnished by the fact that they were joined by the non-initiated girls, organized into their own secret council, the nkilion.

  • 2  Two descriptions of groups of youths in the societies of eastern Africa endowed with age systems h (...)

8It was up to the warriors to watch over the non-initiates; any non-initiate that was deemed to have behaved badly was punished and beaten by the warriors, who even had the right to set fire to a house if they thought a fugitive non-initiate had taken refuge inside. When the official initiations seemed to be too long coming, the fear of the lamalle council—which officially represented the warriors—and of the Fathers of the country, was that the ncibi of the non-initiates would become more powerful than they were themselves. The confrontation crystallized around the warriors’ right to decide whether or not to go and fetch the circumciser. The non-initiates about to be initiated—in line for promotion, as it were—were given a proper noun that was identical for all the territorial sections, anticipating their position in their future set2.

2. The recruitment of male sets

9Each generation set is given a name out of a recurrent series of eight names (fig. 6). When someone is asked to name the list, he will always start with his own generation set. A man from the Michubu set, for example, will continue with Ratanya, Lubeeta, Miriti, go back to the sets immediately before his own, like Ithalie and Kiramunya, and then mark a pause before naming the more distant sets. Sometimes, people cannot remember these latter, and they end up explaining, with the help of gestures, that the old names will return. The idea of cycle and of the return of names that is so clearly expressed comes to the rescue of failing memories, and the list of eight names is the common denominator by which a Tigania-Igembe whole could be defined in former times, and still is defined today. The only exception to this is the Kiramunya set, to which recruitment was in progress at the time of the British conquest, and which was the first set to be promoted to the grade of Fathers of the country under colonial domination; the presence of synonyms, known to the elders, testifies to the disruption suffered by the system at that time, but the reappearance of the cyclical names is also a sign of its resilience. Some people explain that this list has not always been cyclical, that it became so in a very distant era, unlike the Imenti of Mount Kenya, for example, whose list has remained linear. These names, which everyone can anticipate, are attributed, during large public and festive gatherings, to the set that has just attained the grade of political power. Until then, only provisional names are used, in principle.

Fig. 6 — Cycle of names of nthukî, male Tigania-Igembe generation-sets




Kiramunya or Kobai (Tigania); Murungî (Igembe)








Fig. 7 — Rules governing the distribution of sons (from different fathers) in the nthukî. (The generation sets have been numbered for the sake of clarity).

Fig. 7 — Rules governing the distribution of sons (from different fathers) in the nthukî. (The generation sets have been numbered for the sake of clarity).

10At first sight, the rules for joining the sets appear quite simple (fig. 7). All sons must be included in the system, but a son cannot belong to the set adjacent to that of his father; he must join one of the subsequent sets. If a father is in set 1, his son or sons cannot belong to set 2, they can only be in set 3 or after. A father in set 2 cannot have sons in set 3; they can only join set 4 or subsequent sets. A father in set 3 can only have sons in 5 or after, and so on. We therefore have the combination of a positive rule of inclusion, a negative rule introducing a minimum distance between a man and his male offspring, and a practice, that is to say a formula that combines a mechanical principle and a statistical principle.

11By taking the father’s set as the point of reference for defining the position of his sons, the rules of inclusion separate and at the same time connect them within the set system, according to a generational principle that is expressed very simply by the term nthukî which designates these sets. When asked what his nthukî is, a man will give one of the eight names listed above. Another, less-often-heard acceptation specifies the extension of this term. Thus, the number of nthukî between an individual and his forebears is counted to ascertain the genealogical level at which a certain marriage was concluded, and to verify that the payments resulting from marriage transactions have been duly made. The meanings of nthukî cover three complementary but different notions: the generation as a set, the genealogical level, and filiation. This last sense is found in the term that designates the son: as long as a boy is not included in a specific set, he is a mwîjî, a “non-initiate growing up”; his father uses this term when speaking of him, and uses his child-name when addressing him or referring to him. As a result of his integration into a set, the boy becomes a muthaka (pl. nthaka). From that time on, the father uses this term, which entertains a certain ambiguity, sometimes intentional, when a given muthaka is no longer the dashing young “Bushman” to which the term refers, but a man who, in turn, became a father long ago.

12According to another rule that applies in conjunction with the above, only one set at a time can recruit new members. This principle, which underpins all systems of age and generation sets, entails, for example, that if Michubu is being formed, then this is the only set that is open and can incorporate new warriors. The preceding Ithalie has been closed, and the following Ratanya will only open when the Michubu closes. This rule, whereby individuals each wait their turn to form sets with pre-existing names, helps to explain why the elders start with their own set when giving the list of set names, then listing the names of subsequent sets before moving back in time to the sets preceding their own. In this list with no beginning and no end, the recruitment of a new set represents an accession, rather than a starting point, for the individuals that make up the set. However, another condition, of a political nature, must also be fulfilled. The opening of a new set must be authorized by the next-but-one set above it, when this latter set attains the grade of Fathers of the country. Thus, Michubu can only be opened when Kiramunya reaches that grade. In this way, the recruitment of sets is conjugated with the grades described above.

13In summary, six principles govern the constitution of sets and the recruitment of individuals:

  1. The sets follow each another in a pre-established order.

  2. Every man must be included in a set.

  3. A negative rule establishes a minimum gap between a father’s set and his sons’ set or sets.

  4. A principle of a statistical nature ascribes a place to his sons.

  5. Only one set can recruit at a time.

  6. A set opens for recruitment when the next-but-one set above it reaches the grade of political power.

  • 3  Lambert (1945, typescript, p. 409) compares chankanabiri to the Maasai term il-chang’en opir, whic (...)

14Additional rules specify the practical conditions of recruitment. Inclusion in the set follows the circumcision and installation of an individual in the warriors’ dormitory, also called gaaru, but these sessions are not organized continuously. Apart from the fact that they can only be held during thaano, the dry, cool period which is also when the year changes, these lengthy rituals, costly in both time and social energy, are only performed when there is a sufficient number of candidates. On average, the circumcision rituals are (were) held every eight kiingo, according to the six-month calculation in vogue, that is to say an interval of four years, but it can be less and is often more. These rituals supply the framework for the internal organization of the set. Boys cannot be initiated outside of these sessions. Individuals who are circumcised during the same session belong to the same subset, and when three successive subsets have been formed, they make up a set. The subset is called etaana (pl. mataana), a noun derived from the verb gûtaana, meaning “to circumcise” and also “to excise”, just as the word ntaanî, mentioned above, designates the initiate during his period of seclusion. Three ordinal terms recalling their rank in any given set serve as a point of reference for the subsets. The first of these is ndinguri, a noun of which the literal meaning, known to everybody, combines two verbs: kûinga (“to close”) and gwetera (“to wait”). Thus, the ndinguri are “Those who had to wait because it was closed”, and who have therefore grown up and are older than the norm. In certain Tigania sections neighboring the Imenti where the term chankanabiri is in use, this word, when used together with ndinguri, conveys the same idea, since it designates the flowering stem of the sugar-cane which sticks out far above the rest of the plant3. Four or five years after the recruitment of the ndinguri comes the turn of the kobia (the literal meaning of which was not found), who constitute the intermediate subset. Recruited four or five years after the kobia, the kaberia are the last to be initiated into any given set, and the formation of this subset signals the closing of the set. The literal meaning of kaberia is probably “the Little ones who are before” or “in front”: the prefix ka- is a diminutive, and mbere, an adverb of space and time, means “before, in front of”. This composite name refers to the fact that on average, the recruits of the last subset are younger than the ndinguri of the following set; these latter must wait for a new set of Fathers of the country to be promoted to power before they can be initiated, and they end up being older than their predecessors.

15Despite the overlapping of two successive sets that is implied in the literal meaning of kaberia and ndinguri, the subsets within any given set tend to be comprised of individuals of fairly homogeneous and decreasing ages: the ndinguri are older than the kobia, who are older than the kaberia. These divisions resemble age sets that emerge as subsets of generation sets, but when the Meru explain the terms in more detail they emphasize the fact of being initiated together, and the first subset of any set, which brings together men aged thirty or more and younger men, reminds us that the main principle at work here is not the criterion of age. Furthermore, these subsets are identified by proper nouns evoking a particular event or circumstance that marked the session, but over the long run these names are forgotten, whereas the memory of the ordinal terms remains. The historical particularities are at the same time noted and destined to fade away before the principle of repetition of the pre-established structure. With regard to generations, the rules governing the formation of sets mean that every man is identified by two criteria: the name of his set, followed by the rank of his initiation subset. For example: Ithalie kaberia, Michubu ndinguri or Ratanya kobia.

16One last rule, applied insofar as possible, includes the initiate in an additional division, to address the following concern. The rituals are performed in the public ground known as kieni kia gaaru (“the consecrated ground of the warriors’ house”) which every territorial section possesses, but the candidates in one territorial section are not all circumcised at the same time. During one session, three or four groups succeed each other at intervals of several days, following an immutable order: thiana, keenda, mutitimu, muruara, although these terms, recorded at Athiru Gaitî in Igembe, may vary from one place to another. By preference, a boy joins the group in which his father and his father’s father were circumcised. The fourth and last group is only formed if there remain latecomers or boys who live far away. This rule affirms the ideal of filiation already observed in the rules of inclusion, and even specifies a principle of patrifiliation. However, this is a matter of category, rather than set. If possible, a boy is initiated on the same day as his father and his paternal grandfather: this is desirable and favorable, but it is not an obligation.

3. Ntuiko, the handover of power and the progression of male sets

17The basic principles just presented say nothing about the way the whole system moves forward, and the question of how each set of fathers in turn reaches the position in which they authorize the initiation of a new set of warriors remains unanswered. Of course, the ageing and renewal of the population does play a role, but this process operates in every form of social organization, and so it cannot, alone, account for the dynamics specific to the Tigania-Igembe system. Another rule comes into play, which the informants consider to be distinct from those described above because it relates to the political order and its conflicts, in which they are perfectly correct, but this explains precisely in what ways it influences the conditions of recruitment and thereby helps to structure the whole system. The eight sets are distributed alternately into two divisions A and B: if 1 is in A, then 2 is in B, 3 is in A, 4 is in B and so on (fig. 8). Authority is considered to be held by each division in turn, to such an extent that the colonial administration spoke of a two-party system. How are these divisions organized? Each division is a gîtiba identified by a proper noun, Kiruka and Ntiba—the latter term being close to the substantive—in Tigania, Ntangi and Mbainî in Igembe. These divisions are composed of nthukî-generations, and it is in this context that the relations in this order can be most easily read. Thus, A1 are the grandfathers of A5, who are in turn the grandfathers of A1. The repeated cycle of set names is also revealed here, since in each division, the same name recurs in every fifth generation: the set of the grandsons of grandsons takes the name of the set of the grandfathers of grandfathers. One says, for example, that it is the return of the old Gwantai, or that a gîtiba has expired, gîtiba then designating a division, a “cycle” or even a circle (gîchulûûnko) within which one “counts the generations” to determine which one is returning.

Fig. 8 — The dual division of the set system (gîtiba, pl. ntiba)

Fig. 8 — The dual division of the set system (gîtiba, pl. ntiba)

18An additional measure of time appears here, where the cycles of five generations are associated with recurrent, opposing atmospheric conditions: one cycle is associated with the sun and the other with the rain. It is believed that the return to power of the first cycle is accompanied by a period of drought, in the sense that the dry seasons will be more pronounced than the rainy seasons. Conversely, the return to power of the second cycle will coincide with more pronounced rainy seasons. With this multi-year oscillation derived from the empirical observation of wetter and drier episodes, which meteorology is now beginning to explain, the Meru associate ideas of differing prosperity. The gîtiba of the sun is favorable to the multiplication of livestock, humans and bees, but it is also associated with war, whereas the cycle of the rain brings abundant harvests and a return to peace. The same opposition may be expressed in a different way by explaining that when there is a lot of livestock, there are smaller harvests, or that the abundance of livestock is detrimental to the harvests, and vice versa in each case. Even more simply, it is said that one of the gîtiba is associated with the idea of prosperity and the other with the idea of misfortune, but without specifying which is which. These two divisions also relate to the elementary ego-alter dyad (ûni-ûgwe), which covers the same ambivalence of complementarity and rivalry. In these conditions, it is not surprising to see the classifications specific to women also appear in the context of the gîtiba.

19In terms of political organization, the functions are distributed in the following way (fig. 9). At time t, authority is held by A, but in reality it is held by one particular set in the A division. Imagine, for example that set 3 exercises power, as “Fathers of the country”. In that case, set 5 is open for recruitment and is gradually formed with initiates who become warrior- “Bushmen” under the command of the A3s. The A5s may be the eldest sons of the A3s, the younger sons of fathers from the opposite division (the B2s) or the youngest sons of A1s with particularly long paternity. So not all the fathers of the A5s are in the A3 set, since there are some in B2 and even in A1. This results from application of the principles III and IV governing recruitment and the formation of sets. What is happening in the other division at this time? The B4s have left the state of warrior and become young fathers, busy forming their own family: a future B6 set is in the making. The B2s are fathers who have left political affairs behind and who are, one by one, becoming grandfathers. Gradually, with their wives, they carry out the ritual that transforms them into Accomplished. The distribution of sets and functions between the two divisions reveals an important feature of the Tigania-Igembe system: each set corresponds to a grade, and no grade is left empty.

20This equilibrium continues for a certain period of time, and then alters as a result of demographic changes. After a few years, tensions emerge, mainly due to the growing impatience of the B6s, the sons of the B4s, who want to be initiated but cannot do so until their fathers reach the grade of Fathers of the country. The B6s find it more and more difficult to support their state of social minority, but the A3s cling to power and the rivalry between the two divisions intensifies. The A3s seek to initiate the largest possible number of sons in A5 by recruiting in the last subset of sons, who are younger and younger and maybe even too young, in the name of a divisional spirit whereby every father tends to think that only his cycle is beneficial and that the other cycle brings misfortune. For their part, the non-initiated B6s, some of whom may be as old as thirty, become increasingly provocative and pugnacious and end up gaining the upper hand over the official warriors, who are younger than them. This situation is absolutely unacceptable, since, by definition, warriors take precedence over non-initiates, and yet it is designated by the literal meaning of the names for the consecutive subsets of two successive sets, kaberia and ndinguri. As the A3s persist in not relinquishing their position—or the “country”, nthîî, in native parlance—the non-initiates, organized into their clandestine councils, exert additional pressure by threatening to leave with the uninitiated girls and become Kirio (Turkana). The latter are Nilotic pastoralists who live (lived) in the semi-desert regions to the north of the Nyambene Hills, who conducted active bartering and maintained peaceful relations with the Tigania-Igembe. Their society is also organized into generational divisions and age sets, but they practice neither circumcision nor excision. In political terms, this marks the beginning of a period of crisis, of open rebellion instigated by individuals who refuse to be sidelined and who threaten, substantiated by the preparations make, to secede and thereby durably weaken those who remain. The rebels are incongruous individuals, because of the widening gap between their physical age and their condition of non-initiate, and although their rebellion is recurrent, there is nothing ritual about it but status inconsistency (Lenski 1954). The colonial administrators, responsible for maintaining public order, kept close watch on these affairs and wondered about them, but without realizing that this dysfunction is actually the driving force behind the system. The crisis is a property of the structure and ensures its renewal.

Fig. 9 — Interrelation of sets with age grades

Fig. 9 — Interrelation of sets with age grades

Ill. 1 — “Liria. Most influential man among the Embe” [Igembe]

Ill. 1 — “Liria. Most influential man among the Embe” [Igembe]

1892-1893 Rokeby Archives (etching in Chanler 1896: 258).

Ill. 3 — “Chief Natives of the Embe”

Ill. 3 — “Chief Natives of the Embe”

These elders are Igembe Fathers of the country.

1892-1893 Rokeby Archives (etching in Chanler 1896: 258).

Ill. 4 — “Wamsara [Muthaara] elders”

Ill. 4 — “Wamsara [Muthaara] elders”

1892-1893 Rokeby Archives (etching in Chanler 1896: 136).

21The increasing disorder precipitates the transmission of power: the A3s, who are now obviously incapable of maintaining peace and harmony, must retire and make way for the B4s. To put it another way, more in keeping with the spirit of the thing, the turn has come for the Bs to exercise power: in Tigania, Kiruka makes way for Ntiba, and in Igembe, Ntangi makes way for Mbaini. As new Fathers of the country, whose accession to power has finally been brought about by their eldest sons’ rebellion, the B4s immediately open a new set, the B6s, and organize a session of initiation during thaano: as the prolongation of this month allows the course of the stars to be readjusted, so the initiations restore order to individual trajectories and to the progression of the sets. The new initiates are in the ndinguri subset, the literal meaning of which is fully exemplified: this subset of heterogeneous ages contains primarily the eldest sons of B4s and the younger sons of A3s; the closing of A5 means that they can no longer be included in this set, but will obviously be initiated and included in B6. When a new set is opened up to recruitment, all the sets change grade at the same time. In division A, the former warriors become young fathers and start producing a future set; the erstwhile fathers in power prepare to become grandfathers and soon Accomplished. In division B, the B4s who are the new Fathers of the country now lead the councils; their own fathers in B2 gradually pass away and those who are particularly long-lived become ntindiri; the set of warriors B6 grows in time with the rhythm of initiations.

22The memory of the confrontations accompanying and provoking the transmission of power has an informative function (fig. 10). The most eloquent on this issue are the Michubu ndinguri who, in the late 1920s and early 1930s, fought with the Ithalie kaberia to obtain the right to be initiated; I did not meet any Ithalie ndinguri in a fit state to relate anything, and the conditions of recruitment of the Ratanya ndinguri were appreciably different. Individuals of the kobia level know nothing of this issue and even deny that such a thing could exist. As for the Ithalie kaberia, they are not at all loquacious on the matter, which may be explained by their unflattering position of “little ones who are before”, revealing a planned disorder in contradiction with the ideal of harmony conveyed by the cycle of male sets. Nevertheless, the word ntuiko, which designates the transmission of power from one division to the other, is derived from the verb gûtuika, meaning “to break, to interrupt”, clearly highlighting the rupture contained in this transition: one set closes while the other opens. The synonym ûkûûra gaaru or “uprooting of the warriors’ house”, referring to a great ritual where the ex-warriors demolish their house to make way for the new one, expresses a comparable, if slightly different, idea.

Fig. 10 — Transmissions of power and sets involved since the colonial conquest in 1908

Division A

Division B

Approximate closing date

2 – Gichungî


* Kiramunya – 3


4 – Ithalie


Michubu – 5


6 – Ratanya


Lubeeta – 7


8 – Miriti


Gwantai – 1


* Or Kobai, or Murungî. The approximate closing dates correspond to the years when the supremacy of the kaberia was challenged by the future ndinguri, i.e., the phase of ntuiko, which lost of its virulence from the 1950s (at the time of the 6-Ratanya hence the dotted line!) and almost vanished.

23In his report for 1929, the Meru District Commissioner described the agitation that had taken hold of the population:

Disobedience to tribal and government authority was rife among the younger men and the whole system of tribal administration through the agambi or heads of subclans have completely ceased. Moran [the warriors] in defiance of their customs who indulging freely in drinking about instead of working and the elders and headmen were doing nothing to stop this. Stock thieving from both the Meru themselves and settlers along the borders was causing grave unrest. Old men were doing very little to assist their women to cultivate and the young men nothing at all. It was eventually found that the cause of much of this trouble could be traced to the fact that the ntuiko or ceremony of taking over the country by new ages of elders and moran had started in 1927 continues so 1928 and was still going on in 1929. A ntuiko takes place about every 10 years and proceeds clan by clan for many months but it is customary for each clan after a reasonable period to follow up their ntuiko with an other ceremony called ringa ngondu. The period from the start of the ntuiko up to the ringa ngondu ceremony is regarded as a period of unsettlement and in the pre-government days it was then that spears were blooded and deeds of prowess and daring performed by the incoming age of moran. By march 1929, no single clan had yet gone through the ringa ngondu ceremony at which the incoming age of elders, Murungi in this case, should take an oath to observe and administer tribal customs etc. and is struck their moran in their duties, especially making taboos such evils as drinking and stockthieving by young men. Steps were taken immediately to hurry on this ceremony and it was now be completed on most locations. The district has settled down and the new ruling elders have assumed their authority.
dc/mru. 1/2. 1929: 6-7.

24Here, the Commissioner is referring mainly to the situation in Imenti, where the transmission of power takes place more frequently. This ntuiko saw a confrontation between the future Michubu ndinguri and the Ithalie kaberia warriors, and the transmission between Fathers concerned the Gichungî and the Kiramunya (still called Murungî), for whom the Gichungî had to make way. This testimony is interesting because, in addition to the substitutes for warlike activity and the widespread disobedience which the Commissioner observes and explains in his own way, he also records the rather unusual behavior of the Fathers. On their subject, B. Bernardi supplies some further details:

The ntuiko ritual covers a long period during which the country is happily feasting; the elders of the retiring age-class are at liberty, during that period, to do whatever they like; thus one sees them (or hears of them) going around spending time at drinking party or indulging in the sort of behaviour that would be considered shameful at any other time.
Bernardi 1959: 23-24.

[...] all members of the new age-class [of Fathers] enjoy a privileged and sacred status: they tour the country from house to house and are entitled to enter freely. It is a time when many abuses are silently endured by women and house-owners.
Bernardi 1959: 92.

25So the ntuiko is (was) not synonymous with disobedience and confrontations in which the only protagonists are non-initiates and warriors. Married men of the departing set and those of the set acceding to power join in with their own rule-breaking, and the transgression perpetrated by the new Fathers of the country is somewhat reminiscent of the attitude of the circumciser during the initiation. This disturbance of normal behavior takes (took) over all the citizens, and the society is in danger of disintegrating. After such commotion, “clan by clan”, in other words territorial section after territorial section, the ritual of “beating the ram”, during which a ram is sacrificed by the new Fathers of the country, is intended to dispel the effects of these transgressions and restore normality.

  • 4  A. D’Abbadie (1880: 175), writing in the nineteenth century, described the five divisions of the g (...)

26The details of the way the Tigania-Igembe system functions are becoming clearer. The sets are entirely interconnected with the grades. In this configuration, the recruitment of new warriors entails that at the same time, the sets change grade and transmit power from one to the next. However, this translation cannot be carried out smoothly, and it inevitably includes a phase of dissociation between the grades and the sets in order for a new combination to emerge. The ntuiko, a time of widespread disobedience in which normal activity and the application of rules are suspended, is the recurrent means by which the links between grades and sets are dissolved and then reformed. Another interesting aspect is that power is not transferred from a set of fathers to a set of sons, but from a set of ageing fathers to a set of fathers who are younger and in the other division. These generation sets are different from those of the Karimojong cluster to which the Turkana belong, but they are similar to those of the Maasai, and even more like those of the Borana, although not exactly the same. The A and B divisions correspond to fractionated generations, but unlike the Borana, who have five divisions where every son automatically joins the generational division of his father, the Tigania-Igembe divisions, of which there are only two, recruit the younger and youngest sons of the rival division. However, the three or four micro-divisions that appear within each session of initiation delineate a model in the form of an ideal whereby, whatever division they might belong to, the paternal grandfather, father and son all belong to the same unit. Another important feature is that, by virtue of principle III, father and sons are always separated by at least one set, and the inescapable gap between a father and his eldest son, in particular, creates the structural tension that drives the whole system forwards. One division holds authority as long as the other does not display too much impatience, which heightens at the speed of demographic growth and according to the relations between generations that are established within each division. Transmission does not take place after a fixed, pre-defined period of time, as it does among the Borana, where power is automatically transferred to the following division every eight years4; it depends on the balance of numbers between the A and B divisions. This question can only be fully addressed once we have completed the table of sets by considering the position of women.

4. Women in the generation-set system

27A casual observer might conclude that the set system for women is of a secondary nature or even non-existent. The men refuse to talk about these matters, and their silence leads one to think that there is no classification. In reality, their attitude conceals their embarrassment or even fears, because it is unseemly and dangerous for them to speak of women’s affairs. Signs of classification exist in the grade structure, but one must also be prepared to attach sufficient importance to them. The women are incorporated in an unexpected way that sheds light on the system.

28Like the boys, the girls are aligned in relation to their fathers. Whereas it is said of sons that they cannot join the set adjacent to their fathers’ set, it is said of daughters that they cannot marry a man from their fathers’ set. At this first level, a daughter is not recruited into a particular set; she is included in a category of women within which the men of a certain set cannot take a wife. A term of address and reference, ntaliba, with the same root as the word gîtiba, designates for a girl the men of her father’s set and their wives. It signals the avoidance and respect that characterize their relationship. This elementary measure has the double effect of emphasizing the position of a daughter with respect to her father and of raising the question of alliance by means of matrimonial prohibition. The proscription, both matrimonial and sexual, is strict and its transgression is a form of incest, between a “father” and his “daughter”. Any offender is a kioncho punished by exclusion from his set, which amounts to stripping him of all his rights. Departure is the only solution open to the transgressor, although he might hope that a distant territorial section will be prepared to reintegrate him, after a period of waiting and in return for payment of the big fine known as mirongo ithato, or “Thirty”, consisting of heads of bulls, small ruminants and pots of honey. Kioncho also designates a child born prematurely and fated to die. In both cases, this term highlights a disturbance in the order of time and accomplishment, referring to individuals whose development has failed: one comes out of his mother’s womb too soon, and the other is expelled from his class. A different term, win’janga, designates someone who commits incest within the family. The two types of incest do not overlap, nor is one an extension of the other. Among the Igembe, the rule of matrimonial prohibition is accompanied by a positive rule by virtue of which the girls of specific sets are destined to become the wives of certain other sets of men. It is said, for example (see fig. 10), that the daughters of Ithalie are forbidden to the Ithalie and destined to marry the Michubu and Ratanya, or that the daughters of Michubu are destined for the Ratanya and Lubeeta, and so on. In addition to the prohibition concerning the “fathers” and their “daughters”, the Igembe have preferential rules involving the two sets following that of the father. The Tigania only pronounce the proscription between “fathers” and “daughters”, but J. Mahner (1975) has clearly demonstrated that the practice of preferential marriage between these sets nevertheless exists; a situation I have also observed. The Tigania do (did) the same as the Igembe, but without formalizing it.

29The presence of rules of affinal alliance associated with the generation sets is not an isolated fact; they are accompanied by other rules that reveal a direct influence of the system in the sphere of marriage through the political authority. It is within the remit of the council of Fathers to modify or to refuse to modify the rûracio, the part of the bridewealth paid in head of cattle. The ideal—and the practice—is to maintain it at a sufficiently low level to allow every member of the Tigania-Igembe to get married; the reminder or public reaffirmation of the egalitarian norm is issued by the council of Fathers in response to pressure from elders desirous to become polygynous and hoping to marry young women by discreetly paying higher bridewealth. By this means, the alternative between monogamy and polygyny is also established, insofar as the latter is considered to jeopardize the right of every person to get married. Thus, without actually condemning, far less prohibiting polygyny, the council nevertheless exerts (exerted) an influence on the number of marriages into which a man can enter. A different but complementary type of control is exercised when a man gets married for the first time. An individual can embark on this course after having been initiated and completed his time as a warrior, and from the moment when his generation set has been authorized to marry by the Fathers of the country. In principle, this authorization is given when the kobia, the second subset, have been or are being recruited. The ndinguri, the first subset, are no longer the only ones in the warriors’ gaaru to meet the needs of defense; other initiates have been incorporated and there is a sufficient number of warriors; the oldest ones can therefore get married. In principle, the authorization they receive from the Fathers is valid for their whole set. Those who come behind them in the same set can get married when they want or when they can, and the system’s main influence is on the moment the ndinguri get married.

30However, secondary rules also have an effect. Within the set, control is exerted by the older warriors over the younger ones. Under penalty of a good thrashing from his older companions, a junior cannot set about courting a girl until he has “bought the bed” (kûgûra ûruri); the girl must bring tobacco, various foods, butter and ochre to her suitor’s companions. If they agree to receive these gifts, the warriors enter their gaaru, share the food and coat themselves with a mixture of butter and ochre before coming out with the lover. The nkenye and her warrior can now cultivate their relationship in the girl’s house when her parents are absent from the kraal. Courting an nkenye does not always mean marrying her. Some individuals particularly appreciate the warrior’s life and prolong it, while others find it tedious and are impatient to leave it and get married. For this reason, younger brothers, initiated after their elder brothers of course, may find themselves getting married before them; the elders, quietly ironic, explain that such things are inevitable because some warriors have more success with the girls than others. The obligation to win a name during a raid (known as the nthukî name) and the esprit de corps that is so strong among the warriors also tend to delay the decision to get married. A contrast between the mechanical model and the statistical model, already noted elsewhere, can be observed again here: when a prohibition is lifted, the mechanical effect of that prohibition is replaced by norms that have a statistical effect on individuals’ practices and choices.

31In summary, through at least two aspects—the amount of bridewealth and the moment when a set is allowed to marry—the council of Fathers regulates (regulated) the mechanisms that directly influence the reproduction of the set system, and act as a complement to the rules of affinal alliance stipulated with respect to the girls.

  • 5  When answering questions about “belonging to sets” that served as a preliminary to interviews, cer (...)

32Having said that, these various rules are as much a result as a cause of the rivalries between adjacent sets of men. The sons of a set of Fathers affirm that the daughters of that class belong to them, apart from their own sisters, of course, and other relatives forbidden in the universe of kinship in the strict sense of the term. This question will be examined further in the next chapter, but for now, let us just say that the daughters of a set of Fathers develop closer relationships with the set of sons of those Fathers if only because they have probably grown up together in the groups of neighborhood children. The nkenye (uninitiated girls) of a set of Fathers gradually form categories that are the counterpart to those of the uninitiated boys and which foreshadow the subsets of future sets. These relationships become apparent when, for example, the right to form a set is refused for too long; with evidence to back them up, the rebels threaten to leave with the girls they consider to be theirs. Once they have become warriors, the boys continue to maintain their special ties with the daughters of their “fathers”. However, there is an essential ambiguity in the order of the structure, since the warriors cannot immediately marry the girls they consider to be their own. It is therefore inevitable that some of these girls marry men from the next set up, becoming wives in the opposite division by moving up half a generation. Thus, there are several ways of setting out the rules depending on whether one is father or son (fig. 10). An Igembe father of the Ithalie 4 set will say that his daughters are destined firstly for 5 and then for 6. His son will say that they are destined firstly for 6, and any warrior, whichever division he belongs to, will consider that the uninitiated daughters of the class of his Fathers belong to him; a feeling shared by the girls for whom the only good marriage is with a man from the set of their fathers’ sons5. Setting aside the perspective of sets and taking up a more individual point of view, it appears that a father in 4 can have married daughters in 5, 6, 7 and 8 and initiated sons in 6, 7 and 8.

33Girls are (were) initiated long after puberty, and the excision undertaken with a view to their marriage puts an end to the condition of nkenye. The sessions of initiation are organized to the rhythm of the boys’ initiations, in other words during the thaano of the year in which they take place. The same verb kutaana designates the operation, and the initiated girls receive, in parallel to the ordinal male subsets, a collective name that delineates the cohort of initiates, who are also called etaana (pl. mataana). A synonymous term îrika (pl. marika) is known but rarely used, except in the small section known as Thaicho, which lives apart, to the south-east of the Nyambene Hills; rika also designates the grade and the set, and in this case it is equivalent to nthukî. The way that warriors and female initiates are recruited in parallel is an extremely clear indication that the two groups are destined to marry each other, and it prefigures the parallel development of the joint sets of husbands and wives. When a woman marries, she joins the set of wives linked to her husband’s set, as well as the ordinal rank of his subset. Consequently, some women become wives in the set at the level of which they were initiated, while others change level. The otherwise highly interesting summary table of names drawn up by Lambert (1956: 38-39) fails to bring out this shift (fig. 12). The names of the wives’ sets are subject to certain repetitions, but they do not have the same automaticity as those of the men’s sets. There is always a phase of latency during which several names circulate at the same time, about which discussion is rife because some of them are considered more propitious than others. A consensus gradually emerges, and one sole name, proclaimed during the ceremony of naming the sets, is adopted, although synonyms exist between the Igembe and the Tigania. The same ceremonies also witness the proclamation of the proper nouns used by the two sets between which power has just been transferred when they address or refer to each other (fig. 11). A new name is chosen each time power changes hands: “Bamungo” when 2-Gichungî has trans mitted power to 3-Kiramunya, followed by Bantakuû, Baere, Baichû. The oldest names have been forgotten, and it might be possible that a name linking the 6s and the 7s will eventually end up in circulation, given that the 7s are now initiating the new Gwantai, although initiation ritual has undergone tremendous change since Ratanya (see Ch. 10 and Ch. 11). The prefix ba- indicates equality, and the root indicates the thing that has been shared, for example honey (nainchû), as a result of which the 5s and the 6s call each other Baichû. This procedure, which combines gesture and speech, has the effect of putting on an equal footing two sets that had previously been at different levels in the hierarchy. Baite is also used by the men of the two sets between which power has been transmitted to address and refer to each other, and it is used more generally between the men of the two divisions. This is also the term shared between a maternal uncle and his nephew after the latter’s initiation (and the nickname of the Meru people in Nairobi).

Fig. 11 — Women in the generation-set system

Fig. 11 — Women in the generation-set system

Bamungo terms of address between generations that have transmitted power to each other. From Michubu onwards, this has become a way of speaking!

Fig. 12 — Provisional and definitive denomination of the subsets and sets (Male Female) (Lambert 1956: 26, 35-36, 38-39)

Fig. 12 — Provisional and definitive denomination of the subsets and sets (Male Female) (Lambert 1956: 26, 35-36, 38-39)

The nthukî (in small capitals) carried provisional names chosen by the ndinguri (first subset), which often, but not always, corresponded to their initiation names:
– Nkirii, a contraction of Kiraithie, for the Kiramunya;
– Kiambarua, “Those who have written things” (i.e., identity papers), for the Ithalie;
– Meme for the Michubu
The table of provisional names is in fact less homogeneous than presented by Lambert. Variations existed from place to place but equivalences were easily established. The disruption of female initiations at the end of the 1930s explains the disappearance of names from Michubu onwards.

Ill. 11 — Young Mothers and young Fathers dancing together (Igembe)

Ill. 11 — Young Mothers and young Fathers dancing together (Igembe)

1892-1893 Rokeby Archives.

34The sets of wives advance jointly with their husbands’ sets, sometimes in the company of the latter, sometimes separately. In addition to a daily sociality highly influenced by these divisions, the women accede with their sets to positions of authority marked by the meeting of councils specific to them (kiama kia aka) and possession of the kaaria, before exercising the ritual functions of the mwariki. Without entering into too much detail here, two aspects of the position of women in the system should be mentioned. Firstly, there exists, between husbands and wives of the same set, an institutionalized relationship of lovers, known as bankiro—perpendicular, in a way, to the ntaliba relationship marked by strict sexual and matrimonial prohibition—which every husband, himself involved in this game, may either encourage or frown upon. Both cases occur. If a crime is committed within this relationship, the council of Fathers, which usually organizes the payment of the blood money, does not intervene because it is said to be an affair internal to the nthukî, not concerning anyone other than the members of that set; no payment is made except for the sheep destined to eliminate the pollution of death. Any child born of a bankiro union is considered to be the child of the official husband, who is, in any case, in the same generation set as the lover. Secondly, the situation of the young wives of polygynous men reveals underlying tensions that can be instructive. A young wife belongs to the set of wives conjoined to her husband’s set, but she follows the life and rituals of the set of wives of her own age. Her set of reference will have already moved up through the grades; it would be unthinkable for a woman not to traverse them herself, and in particular not to possess and then transmit the kaaria at some time in her life. Although this type of arrangement is frequent in generation-set systems—like the situation of late-born sons among the Borana, it is nevertheless the result of a mismatch between the age of the woman and her position in the set system. Both in the set to which she officially belongs but in which she does nothing and in the set in which she is active but to which she does not belong, the young woman finds herself in an awkward position, difficult to come to terms with. This observation and the gradually-discovered practices that counteract the practice of polygyny (control of the rûracio by the council of Fathers, the rules of affinal alliance and the attitude of the warriors who watch over the girls that they consider to be theirs by right) combine to make attractive the ideal of marriage between individuals in the same set position, and to reinforce the practice. In this respect, lastly, the ritual complementarity between the class of wives who own the kaaria and the class of warriors—very often their own sons—also results from a convergence of interests between sets, admittedly in different positions, but which nevertheless both work against the polygynous aspirations of the fathers.

35Let us now summarize the principles governing the constitution of female sets. They are called nthukî, like the male sets. But the men affirm that these are not real nthukî, in which they are right, because the women are not included in them on the grounds of their filiation, which is the characteristic feature of male nthukî; they are not born into their nthukî, which the men consider to be their own situation, but they join that of their husbands. However, this should not lead one to conclude that these classes, only vaguely resembling those of the men, serve no purpose—the judgment suggested by the rare comments the men allow themselves. In reality, the female sets are the product of a combination of filiation and above all affinal alliance. Firstly categorized as the daughters of X, explicitly destined in a division for Y and Z, the women are then organized into successive sets as wives. The relative position of wives’ sets and the ties that the women form within the sets organize the women amongst themselves and the relations that develop between men and women over the course of their lives. The structuring effect of the Tigania-Igembe age and generation system appears to be pushed to the limit. Now it is easier to understand why men and women have a different image of age sets. Because of the path they have followed, the old women generally have better knowledge of the whole: categorized as the daughters of a given set of fathers, who must avoid the men of that set but who can approach the men of younger generations, they experience very strongly the ambiguities of the system through its matrimonial rules, and discover the other side of the mirror, as it were, from where the men are so busy. Next, they constitute a set of wives belonging in a certain division, finding their place among other sets of wives and men. As for the men, once they have been initiated they no longer change set, and they know the sets in close proximity to their own: they can only incompletely list the cycle of names presented at the beginning of this chapter (and which the women do not express), they can only classify in their own divisions those sets that are close to their own, and nothing more. The feeling of belonging to their division of generation takes precedence over knowledge of the internal composition and general structure, to such an extent that the only good and true division is his own, whereas the women emphasize the complementarity and relationships between and within the different divisions.

Ill. 12 — Mothers and Fathers getting ready to dance (Igembe)

Ill. 12 — Mothers and Fathers getting ready to dance (Igembe)

1892-1893 Rokeby Archives.

Ill. 13a & b — Mothers and Fathers dancing together (Igembe)

Ill. 13a & b — Mothers and Fathers dancing together (Igembe)

1892-1893 Rokeby Archives.

5. The dynamics of the system and control of demographic drift

  • 6  I discovered this crisis in the colonial archives (DC/MRU. 3/7. 1947). When questioned on the subj (...)

36Now that we have the constituent elements of the generation-set system, we can start thinking about its overall dynamics. The system is driven by the rules of recruitment and inclusion of individuals and the rhythm of the opening and closing of sets, which function as so many time constraints. The recruitment of a new set follows a transmission of power, which is itself the result of rivalries that oppose consecutive sets and divisions; that is, a balance of power that depends on the number of individuals present and the age of the protagonists, in sum, the demographic profile of the society. Apart from the fights started by non-initiates older than the official warriors, the demographic profile can be appreciated through the population of the different buildings related to the generation system, which the Fathers visited and monitored. These buildings allow one to conduct a rudimentary survey of the non-initiates grouped together in the little gaaru at the level of each neighborhood, on the one hand, and of the rising or falling numbers of warriors in the big gaaru, one of which can be found in every territorial section. The Fathers in power seek to initiate the largest possible number of sons in their division, but they are also anxious to preserve a certain demographic equilibrium between the two divisions. A recruitment crisis which arose in the aftermath of the Second World War, when the system had already been greatly weakened by colonization, is very revealing about the issues at stake, as if the declining system allowed certain fundamental principles to come to the surface6.

37In 1946-1947, the time came for the A3-Kiramunya to transmit power to the B4-Ithalie (see fig. 10). The transmission of authority had increasingly little relation to the traditional criteria, but it nevertheless stirred up an effervescence that the administration had to keep in check. The unrest that was the rule during these periods of ntuiko took an unusual turn in the arguments put forward by the A3 to delay the event; the Kiramunya explained that many of their sons A5 (Michubu) had gone to war and never come back. As a consequence, their set was low in numbers, and these losses had to be made up for by recruiting the newly-circumcised men into A5, and not into B6. The elders accepted this argument and agreed to the extension of recruitment, but things did not turn out as expected. Unrest grew: the B6 above all disagreed with this move and refused to be recruited into A5, thereby delaying the holding of initiation ceremonies—the opposite to the usual situation where the candidates tend to precipitate events. The administration had to call meetings to resolve this deadlock, and the elders once again subscribed to the argument advanced by the A3. But repeated intervention by the elders failed to overcome the resistance of the B6, sons of B4, who categorically refused to be considered as sons of A3, by denying the official circumciser of the A division the right to operate on them (as the reader may have inferred, each division has its own circumciser). In the end, the B6 prevailed: the A5 (Michubu) set was closed and the B6 set (Ratanya) was opened.

38This crisis is instructive in several ways. Firstly, it reveals the concern to avoid “empty sets”. The arguments put forward by the A3 were initially accepted because they were technically applicable, because of a rule that permits, under very rare circumstances, infringement of rule III of the system. Those of my informants, of the Ithalie set, who remembered the affair very well for having taken part in the reconciliation meetings, explained that during certain periods when war, disease or famine had decimated a set or subset of warriors, one sought to restore the balance by allowing the Fathers to continue recruiting into the generation set of their sons: a set with an insufficient number of members might induce a long-lasting asymmetry between the generational divisions and result in a paucity of warriors. These arguments illustrate the two concerns motivating those who ran the system during the classic regime: military imperatives and maintaining an equilibrium between the two divisions. By seeking to make up for the losses in their sons’ set, the Fathers were defending not only their sons’ set but also their generational division and therefore the balance between the two divisions and the reproduction of the system. This affair also shows that the system was managed on an ad hoc basis, according to the population numbers that the Tigania-Igembe had the means to estimate. However, we should not conclude that the problem of demographic equilibrium is in fact easy to solve and simply calls for the rules to be manipulated to suit the needs of the moment.

39Secondly, the obstinate refusal of the B6 to be considered as A5 and the arguments they put forward are equally revealing. They explained that they did not want to be governed by “overaged” Fathers, “circumcised a very long time ago”—the A3 were indeed circumcised between approximately 1900 and 1915— or to belong to an A5 set of which many members were already married fathers. This criticism, so clearly expressed, of the mixing and confusing of different ages, reflects a strong tendency of the Tigania-Igembe social organization: there is a time for each thing, and the individual accomplishes it at the right time as part of his set, in which the heterogeneity of ages should be limited. The B6 added that the elders’ argument that a numerical equilibrium should be maintained for the purposes of defense was no longer valid because, since the colonial conquest, the obligations of traditional warfare no longer existed. The disappearance of that imperative brings to light deeper motives: the B6’s refusal to become A5 is fuelled by the “spirit of division”, according to which it is not auspicious to get recruited to the other gîtiba, because the latter brings bad luck. The B6 argued that the activity of the A division had coincided with tragic events: the recruitment of the A3 was contemporaneous with the arrival of the English and the First World War, and the recruitment of their sons the A5-Michubu saw the outbreak of the Second World War. The refusal to be initiated into the other division is also a negative expression of the desire to be initiated in the right position, the position that every man should hold in relation to his father. All in all, the reaction of the B6 reveals rationales that counterbalance the pragmatic nature of the arguments used by the A3 and help to demonstrate that the system may have been run on an ad hoc basis, with adaptations being made according to need, it nevertheless remained true that other factors drove towards application of the rules: over the long run, the system perpetuated itself according to these rules and the values they conveyed.

40Thirdly, the final episodes of the crisis also provide a wealth of insights. Certain A3 fathers persisted in circumcising sons using the A division’s circumciser, in breach of rule V, which stipulates that only one set can recruit at a time. This introduced a decisive break, since it meant the simultaneous opening of two sets, in other words the end of alternation and of the interrelation between set and grade. The administrator who reported this event made the following comment:

[...] as regards the circumcision ceremonies in Tigania. I have no doubt that also the trouble has subsided, the Kiramunya [a3] are not satisfied with the decision and may have circumcized some youths into the Michubu age-grade [a5] but the number is so small that when the main circumcisions take place into the new age-grade as yet unamed [the future b6 Ratanya] the recently circumsized youths into Michubu will of their own accord elect to go into the new age-grade.
dc/mru. 3/7. 1947: 2.

41The administrator can be forgiven for confusing grade and set. He ran little risk of being wrong when he predicted that the few boys who had been circumcised by the Kiramunya would join the newly-opened set. The B6 had the means to enforce this rule because during these years the power of the Fathers as a set had been weakened too much for them to stand up to the “warriors” who felt they were in the right and who possessed, moreover, the wherewithal to exert physical pressure on those who might be recalcitrant. This episode testifies to the resilience of the rule of set-closing and alternation, which survives to this day (1993) -- although it only concerns categories now since interrelation of sets with age-grades as well as grades have vanished-- and underlines the need to distinguish between what relates to the time of the structure and what relates to historical time. One cannot hope to understand the system without disentangling the different temporalities at work inside it. Lastly, this crisis is of a broader comparative interest, since it allows us to imagine the conditions of transformation of one system into another: the systems with two cycles of the Embu-Mbeere who followed the Meru on Mount Kenya, of the Kuria-Zanaki and others, who recruit in parallel and whose age-sets are not linked to the age-grades, present certain indications that suggest that this situation is a result of the abandonment of the rule of alternation between two divisions, a modification that certain Tigania were prepared to accept.

42There are other types of dysfunction that may lead to the system being abandoned. The impatience displayed by candidates waiting for initiation obliges the Fathers to transmit power and open a new set, but this pressure must not be exerted too early. If they were initiated sooner, these sons would become fathers sooner, and in turn produce a cohort of boys who would press to be initiated. The result would be an acceleration in the renewal of generations, and this would prevent the synchronous progression of age-sets through the grades and bring an end to the link between set and grade. This distortion started to appear between the two World Wars, from the moment the colonial administration, anxious to avoid disorder and to produce a workforce, shortened the period of recruitment and “military service” of the Ithalie, who became fathers earlier and produced a larger number of children—the future Ratanya. Conversely, insufficient numbers can also have a disruptive effect, because of the following alternatives. Either the Fathers wait until there are enough warriors in the set of sons, to such an extent that they contravene rule III, or the non-initiates of the other division apply this rule and wait for their set to be opened. In the second case, they are relatively old when they become warriors and then fathers, with the result that they produce less-numerous cohorts, which in turn extends the periods between sets. It is then a general slowing down in the renewal of generations that makes the system unmanageable, leading to the rules and principles being modified or abandoned. Oral traditions, well-known to the elderly, recount that a very long time ago, in the territorial section known as Mûciî Mûkûrû (“Old Establishment”), situated in the Tigania highlands, people got married so late that they could no longer have children, to such a point that the population was heading for extinction. The Chuka, another section on Mount Kenya, had a similar reputation. These two examples illustrate the fact that people are aware of the possible consequences of delaying marriage for too long, and the need to adjust the rules accordingly.

43To summarize, the impatience of the sons waiting for initiation must not manifest itself either too early or too late; both the “full classes” and the “empty classes” of the demographers’ age pyramids render the system unworkable. Consequently, we must either postulate the presence of mechanisms and practices that regulate the flow of generations in one way or another, or we must conclude that this mode of organization can only have an ephemeral existence.

44This problem relates to the now-classic question of the demographic drift that affects all generation-based systems and which can be defined in the following terms. Men of the same generation do not all procreate at the same time; as a result, their sons are of very different ages and their grandsons even more so. If nothing is done to correct this increasing dispersion of ages over time, it becomes impossible to obtain an effective organization into generation-based groups, and yet this is precisely what many eastern African populations have done. For the generations not to be simple categories, but an effective principle for the division of society and the allocation of power, it is necessary for the dispersion of ages and the resulting ever-increasing overlap between generations to be limited in one way or another. The phenomena of “over-ageing” caused by individuals being conceived too early and “under-ageing” caused by individuals being conceived too late have only recently been the subject of sociological study (Tornay 1989b, Peatrik 1995a). The rules and practices aimed at limited demographic drift—which many ethnologists considered to be irremediable—do exist, but they are not immediately perceptible. In this respect, they can be considered as secondary or underlying rules, which does not mean they are of no importance, but that they are only accessible through ethnological study and analysis. The question of how aware the populations are of these phenomena is a dimension that will be analyzed in the last part of this book. For the time being, let us just observe that the Tigania-Igembe have unquestionably endowed themselves with the means to assess the state of their system—the different buildings—and that they implement a set of practices that result in a control over fertility that only makes sense when it is related to the generational system.

45A first underlying rule lies in the domain of the family: as soon as a son or daughter is (was) initiated, their parents must stop having children. The father can take a new wife, but as we have seen, although polygyny is not forbidden, it is restricted in various ways, and so this prohibition affects a large number of men. This rule derives from a particular conception of the person, described in the previous chapter, which here assumes additional significance: the reproductive activity of an individual is deemed to be dangerous for the individuals of the following generation, and consecutive generations cannot reproduce at the same time. Procreation is a right that is received, exercised during a certain length of time and then handed on. That is why the moment of initiation of the children— boys and girls—is postponed as long as possible, since this rite marks their entry into a phase heralding marriage and authorized procreation. The children who are initiated and married at a relatively late age will in turn tend to delay the initiation of their own children, and so on. This constraint, added to the low level of polygyny, helps to impose a statistical limit on the overlap between generations and the phenomena of over- and under-ageing, i.e., the demographic drift caused by children being conceived too early or too late. In practice, the distribution of individuals for whom membership of a generation has been carefully noted can be seen to follow the rules set out above (fig. 7): the sons of a father in (1) are in (3) and (4). The exceptions are known, and readily explained when one asks what the causes are. Apart from the situations of crisis already described, in the event of a father’s untimely death, it is possible for the son of a member of (1) to be included in (2), which allows him to be the deceased’s heir. But for this to take place, two conditions must be satisfied: the son must be above a certain age, since he is supposed to go and fight as a warrior, and the candidate must offer a bull to his deceased father’s companions of generation. By accepting the bull, the latter indicate that they are willing to lose a son to the benefit of the fathers in (2). They share the carcass in an unusual way. One of the rear legs is cut in half crosswise: the lower part, less appreciated, is allotted to the spokesman of the nthukî who have accepted to lose a son. Apart from this dispensation, the sons of a father in (1) are never found before (3). We have seen how the desire to be initiated into the right division also helps to enforce this gap.

46Conversely, we should also examine the maximum dispersion of a man’s children. There is no prohibition imposing any explicit limit, but the records show that dispersion is nevertheless restricted. It may happen that a father in (1) has sons up to (5); in this case they are the children of a polygynous man or a man who had sons very late. The individual included in a set far-removed from that of his father tends to conceal this fact, and only reluctantly acknowledges that his father is a “grandfather”. He will start by declaring that his father is in (3), and only detailed knowledge of families and individuals allow to discover the truth behind these fictions that everyone is prepared to swallow. The silent disapproval of a man who behaves like a father when he is in fact a grandfather, and the discomfort of his youngest sons are indicative of the model that imposes a time limit on the right to procreate, and which is recalled by the preconditions for the rite of the Accomplished, whereby the candidate must, in principle, have initiated all his children. However, in statistical terms, these underaged people pose less of a problem, because they are simply shifted into a generation set fitting to their age. Once that is done, they will, when their turn comes, produce children at the appropriate moment, and will not contribute to the demographic drift.

47Among the Tigania-Igembe, as in other populations that use the generational criterion as a basis of organization, the real problem is not that of individuals conceived too late, but of those conceived too early, who can rapidly cause the system to collapse. To be able to distinguish statistically between successive generation-sets, the men must not embark on their careers as fathers too early. The Tigania-Igembe apply this principle, but they nevertheless raise a formidable problem, since the men who are initiated late are placed in a contradictory situation. They have the right to enjoy sexual relations with the women in the category of their potential wives, but they must under no circumstances engender children. But this can hardly be avoided, despite the techniques of coitus interruptus they practice. Here, the dread aroused by the monstrous creature kitheega conceived by an uninitiated woman takes on its full significance; especially since in the native gloss, it is also justified by explaining that it would be impossible to initiate old boys and young boys at the same time.

48All in all, the abortion or infanticide of babies conceived too early, late marriages stemming from the prohibition of simultaneous procreation by two different sets and upheld by the rule that the marriage of the first subset of warriors must be authorized by the Fathers, result in the statistical maintenance of a gap between the moment when a man starts to procreate and the moment when his eldest son has the right to do so: on average thirty years separate these two events. Rule III can be applied. The Tigania-Igembe nthukî are not age sets or generational fictions, and certain “surprising” practices can only be explained with reference to a generation-based order. The generations are simply fractionated: a father of generation (1) sees his older sons recruited into (3) and the younger ones into (4), according to their order of birth and the rhythm of the opening and closing of sets. The principle of alternation reduces the rivalry between divisions but it is as much a product of their equivalence and contributes to their maintenance; consequently, this dual division is a secondary rule that plays a part in the constitution of sets.

49The principles of recruitment of women and the system’s rules of alliance can also be analyzed in terms of the control of demographic drift, and their initial raison d’être may reside therein. The matrimonial categories and the prohibitions that accompany them regulate the age at which women get married and limit the age difference between husband and wife. In demographic terms, this is an important mechanism that favors monogamy and thereby limits drift: in every population where the numbers of men and women are roughly equal, polygyny is all the more likely to occur when the age difference between men and women at marriage is large, which means that the young men remain single and only have late access to marriage. This leads us to consider a distinct but complementary aspect of this issue: the later a marriage, the less childbearing potential the woman has, and the fewer children will be born. Now, in addition to the late marriage of men and women, the Tigania-Igembe also practice a long gap between each child, since the children are (were) not weaned until the age of five. One could imagine this interval between two children as the equivalent of the interval between two subsets, as if the spacing between births echoed the rhythm of the constitution of sets.

50All these rules and practices combine to maintain a low fertility rate. Moreover, the low number of children per woman, which had, in the past, been so characteristic of the Meru that it was one of the aspects of the conflict in which they opposed the colonial powers (see Ch. 10 and 11), was also promoted by the representations and discourse that condemn large numbers of offspring: they are a sign of sexual activity that is ill-managed by the parents, who behave and multiply like animals—ants or chickens, for example. Once again, the regulation of growth and fertility at the stable level of a stationary population reflects the necessities inherent to the generation-based system: it limits the effects of hollow or full classes that would throw the succession of generations into chaos; it maintains a certain numerical equilibrium between the sets and the divisions; it provides the means to have representatives in every grade.

51Just as the way human longevity is written into the grade system called for comment, so the question of the demographic viability of such a regime of fertility must be addressed; it presupposes a relatively low mortality rate and/or positive net migration. Indeed, it is difficult to relegate these practices of low fertility to the domain of ideology or pure fiction because they help to explain how the Tigania-Igembe sets are (were) formed and how they manage (managed) to bring together individuals of roughly the same age. The nthukî are generation sets for which demographic drift is well controlled, which is also attested by the absence of representatives from the very old generations. The interlinking of sets and grades that this control allows, and of which the stakes are the organization of power and the affirmation of sovereignty, has the effect of making the state of the system dependent on the trajectory of individuals, and vice versa: the system is no more than the sum of the parts. Now it would be interesting to examine the details of this interdependence from the perspective of the domestic unit and kinship, to better understand the mechanisms of institutions that are unusual, to say the least.


1  Even the English word “set” is a source of misunderstanding. This polysemous word conveys various meanings, having different bearing on sociological and political dynamics, that a French speaker would distinguish between “série”, “catégorie” and “classe.”

2  Two descriptions of groups of youths in the societies of eastern Africa endowed with age systems have been written by authors working with very different contexts and approaches. C. Turnbull [1972] 1984: 111-116), who must be complemented by B. Heine (1985), describes the self-distribution of power and food by children among the Ik people, who were experiencing a tragic situation. U. Almagor (1978b), who tested S. Eisenstadt’s hypotheses on the origin of age sets, examined the formation and destiny of bands of youths among the Dassanetch.

3  Lambert (1945, typescript, p. 409) compares chankanabiri to the Maasai term il-chang’en opir, which designates the first subset of each circumcision set. This term, also be found among the Pokot and the Nandi, refers to the large feathers of the ostrich.

4  A. D’Abbadie (1880: 175), writing in the nineteenth century, described the five divisions of the gaada, which he called “castes”, in the following terms: “Of this division into five [...] the Oromo say that it derives from the five natural governments: 1° that of men or reason; 2° that of running water or progress; 3° that of the sheep or tranquility; 4° that of the lion representing strength; and lastly, 5° the government of the vulture, which presides over plunder. The descendants of Maca believe that each caste, when it accedes to power, imbues all its dealings with its own specific tendencies. Thus, the bird of prey inclines people’s minds towards war, whence the dictum Dulo dula, meaning: the government of the Dulo goes on war parties. In this strange constitution, each of the ten sets (of fathers and sons grouped into five fractions) must hold power for eight years, they must each govern in turn, and the sons can only enter the set corresponding to that of his father, that is to say the one with the same tendencies. It follows that a son can only replace his father thirty-two years after the latter has left power.” (Translated from the French).

5  When answering questions about “belonging to sets” that served as a preliminary to interviews, certain old informants, widows of polygynous Kiramunya men, took care to specify that that they should have married Ithalie men, since they had been initiated during the same period as them. Their resentment and chagrin was still alive several dozen years after the obligation that had been imposed on them. Similar feelings could be observed in the wives of other sets.

6  I discovered this crisis in the colonial archives (DC/MRU. 3/7. 1947). When questioned on the subject, some of my informants remembered it particularly well because they had been consulted by the administration, and they willingly added further information. The way this affair came to light is a good illustration of the technique of recall sometimes employed. Amnesia may be due to the passage of time, but in the present case, it is also the consequence of denial that conflict between sets can exist. The affirmation of a neo-traditional tribal council, the njuri ncheke, seeking (and succeeding to a certain extent) to persuade people that its mode of functioning—selective and no longer collective—has always been the rule governing the recruitment to councils of men (see Ch. 10), also played a part.

Table des illustrations

Titre Ill. 8 — “Embe [Igembe] shepherds”
Crédits 1892-1893 Rokeby Archives (etching in Chanler 1896: 250).
Fichier image/jpeg, 174k
Titre Ill. 9 — Boys
Crédits Kianjai section, 1893-1894 (photo by G. Kolb, in Neumann 1898: 65).
Fichier image/jpeg, 97k
Titre Fig. 7 — Rules governing the distribution of sons (from different fathers) in the nthukî. (The generation sets have been numbered for the sake of clarity).
Fichier image/jpeg, 12k
Titre Fig. 8 — The dual division of the set system (gîtiba, pl. ntiba)
Fichier image/jpeg, 19k
Titre Fig. 9 — Interrelation of sets with age grades
Fichier image/jpeg, 19k
Titre Ill. 1 — “Liria. Most influential man among the Embe” [Igembe]
Crédits 1892-1893 Rokeby Archives (etching in Chanler 1896: 258).
Fichier image/jpeg, 157k
Titre Ill. 3 — “Chief Natives of the Embe”
Légende These elders are Igembe Fathers of the country.
Crédits 1892-1893 Rokeby Archives (etching in Chanler 1896: 258).
Fichier image/jpeg, 264k
Titre Ill. 4 — “Wamsara [Muthaara] elders”
Crédits 1892-1893 Rokeby Archives (etching in Chanler 1896: 136).
Fichier image/jpeg, 188k
Titre Fig. 11 — Women in the generation-set system
Légende Bamungo terms of address between generations that have transmitted power to each other. From Michubu onwards, this has become a way of speaking!
Fichier image/jpeg, 31k
Titre Fig. 12 — Provisional and definitive denomination of the subsets and sets (Male Female) (Lambert 1956: 26, 35-36, 38-39)
Légende The nthukî (in small capitals) carried provisional names chosen by the ndinguri (first subset), which often, but not always, corresponded to their initiation names:– Nkirii, a contraction of Kiraithie, for the Kiramunya;– Kiambarua, “Those who have written things” (i.e., identity papers), for the Ithalie;– Meme for the MichubuThe table of provisional names is in fact less homogeneous than presented by Lambert. Variations existed from place to place but equivalences were easily established. The disruption of female initiations at the end of the 1930s explains the disappearance of names from Michubu onwards.
Fichier image/jpeg, 59k
Titre Ill. 11 — Young Mothers and young Fathers dancing together (Igembe)
Crédits 1892-1893 Rokeby Archives.
Fichier image/jpeg, 215k
Titre Ill. 12 — Mothers and Fathers getting ready to dance (Igembe)
Crédits 1892-1893 Rokeby Archives.
Fichier image/jpeg, 153k
Titre Ill. 13a & b — Mothers and Fathers dancing together (Igembe)
Crédits 1892-1893 Rokeby Archives.
Fichier image/jpeg, 312k

© Africae, 2020

Licence OpenEdition Books


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search